Age-specific usability issues of software interfaces

Simone Wirtz, Eva-Maria Jakobs, Martina Ziefle

Simone Wirtz, Human Technology Centre, RWTH Aachen University

Eva-Maria Jakobs, Textlinguistics/ Technical Communication, RWTH Aachen University
Martina Ziefle, Human Technology Centre, RWTH Aachen University

The aim of the present study was to identify shortcomings of electronic interface designs for older
users. Beyond general ergonomic factors, the study focused primarily on visual and language aspects
of interface design. By means of a multi-method approach, combining cognitive walkthrough proce-
dures with a usability test, age-specific problems and requirements of older computer users were ex-
amined. In the cognitive walkthrough shortcomings of a sample application were analyzed. In the
usability test the navigation performance of older users (aged 55+) was contrasted to a young adult
group, to identify usability problems, which are age-exclusive (interaction difficulties only for older
users) and age-specific (general problems that are more pronounced in the older group).

BACKGROUND
In many western societies the group of older adults is one of
the most growing user groups of information and communica-
tion technologies. The rate of computer usage among people
aged 55+ is still low compared to younger age groups, but is
continuously increasing: In Germany, for example, 31% of
users older than 55 years reported to use a computer regularly
in 2004. Until 2007, this rate grew up to 41% equaling an
increase rate of 10%. In the same period, increase rates in
other age groups did were considerably lower (age group
10-24 growth of 2%; age group 25-54 growth of 7%) (destatis,
2008). Furthermore, the usage of new information and com-
munication technologies, like computer technologies, becomes
more and more important and necessary in the everyday life of
older adults (e.g. using ticket machines, doing one’s taxes).
Age-specific usability demands for software interfaces are
examined from different disciplinary and methodological per-
spectives, such as psychology, computer science, economic
studies and engineering (e.g. Czaja & Lee, 2003; Schneider,
Schreiber, Wilkes, Grandt, & Schlick, 2008; Zajicek, 2001;
Ziefle & Bay, 2005; 2006; 2008). Depending on the research
perspective, different issues are focussed on, ranging from
performance and acceptance issues over software features, to
technical parameters. Characteristically, outcomes remain
within the respective research community, mostly not satis-
factorily communicated across disciplines. In addition, hardly
any usability study explicitly focus on language issues, even
though the successful interaction of humans with technical
devices depends to a large extent on the quality of the device
communication. Language issues and technical communication
is an integral part of electronic interfaces.
The design of software interfaces is interdisciplinary by
nature, but there is a lack of multidisciplinary approaches
modeling the relationship between influencing factors as well
as the relationship and interplay of interface elements (design,
language, operational elements, structure etc.). This lack is not
comprehensible and not acceptable facing the increasing num-
ber of old and frail computer users on the one hand and the
solid (disciplinary) knowledge concerning the aging impact for
computer usage on the other hand. Findings from recent stud-
ies indicate that a huge set of different influencing factors has
to be considered (Jakobs, Lehnen, & Ziefle, 2008).
The study at hand combines approaches and findings from
different disciplines to examine the usability of software inter-
faces especially for older adults, to identify relevant language
aspects of software interfaces, and to check the effect of se-
lected influencing factors. Important influencing factors are e.g.
biological or socio-cultural factors.
Biological factors
Ageing is accompanied by a number of physical processes,
which are relevant for the successful interacting of older adults
with technical systems.
Sensorial changes. All sense organs decrease with increas-
ing age. Among the visual functions, visual acuity decreases
color perception and contrast sensitivity decrease while the
sensitivity and susceptibility to glare increases (e.g. Kline &
Scialfa, 1997; Oetjen & Ziefle, 2007; Schlick, Ziefle, Park &
Luczak, 2008). Concerning the design of software interfaces,
changes in the visual system should be taken into considera-
tion for layout decisions, i.e. the choice of colours, contrast,
font, type size, size and design of objects such as buttons.
Furthermore, the ability to hear often decreases in older age.
Older people have difficulties in perceiving soft tones at high
frequencies and when distracting background sounds occur.
Acoustic impairments are relevant for audio elements of soft-
ware applications such as alerts, voice output, and acoustic
feedback. Tactile changes mostly occur at the age of 70. The
ability to distinguish tactile stimuli can diminish and as a con-
sequence movements that demand fine motor skills may cause
problems. Using touch screens therefore can be very difficult
for older people.
Physical changes have an influence on e.g. mobility, (mus-
cular) strength and fine motor skills (e.g. Vercryssen, 1997).
Fine motor skills can be affected by shiver, immovability, and
decreasing sense of touch in hands and fingers. Many older
people have difficulties to operate with input devices such as
mouse or touch pad. Small operating elements at close quarters
can only be handled with great difficulties. Functional deficits
in this vein are therefore relevant for the design of operating
elements (size, type and spaces between elements).
Cognitive changes affect various aspects such as intelli-
gence, information processing, and memory. Within cognitive
abilities, fluid components are distinguished from crystallized
intelligence components. Fluid components include reasoning
abilities, abstract thinking and problem solving processes.
Fluid intelligence is assumed to be independently of learning
processes. In contrast, crystallized intelligence components
refer to accumulated (domain)knowledge, expertise and expe-
rience, including socially mediated, culturally determined
cognitive strategies and skills, such as speaking, reading, writ-
ing, as well as professional skills. Fluid intelligence tends to be
negatively affected by age, whereas crystallized intelligence
remains or even improves (e.g. Baltes, 1991). Crystallized
abilities are strongly linked with socio-biographic (e.g. educa-
tion and educational opportunities) and social factors (techni-
cal socialization). Therefore biological factors cannot be
regarded isolated. Different factors, which have an influence
on computer usage can be closely related and interact.
Moreover, information processing can be affected in old
age. The amount of information that can be processed at the
same time decreases and more time for information processing
is needed (Park & Schwarz, 1999). Acquiring, linking and
memorizing new information are hindered. Searching and
activating memorized information are affected. These changes
should be considered for the development of menus and struc-
tures as well as for the complexity and volume of content.
Working memory declines as well as spatial memory in
older age. Older people have difficulties to localize objects and
to remember non-verbal elements and tend to have poorer
memory for spatial tasks (Hawthorn 2000). These aspects
could be relevant for the design of labels, headings, perma-
nence of important information and objects, the structure and
visualization of content and functions in software interfaces.
These biological factors are samples for changes in older
age. Age-related functional deficits and restrictions may vary
inter- and intraindividually, but, when they occur, they affect
the computer performance, and thus, have to be considered to
guarantee usable interfaces for older adults.
Socio-cultural factors
From a socio-cultural perspective it is important to examine
which kind of technologies dominated during the adolescence
of a generation. Based on Mannheim’s generation theory
(1964), generations have to be understood as groups of cohorts
that differ due to social change. The groups can develop a
collective consciousness because of drastic social occurrences
(Sackmann & Weymann 1994). The term “generation” refers
to contemporary-historic generations, who developed a gen-
erational consciousness because of shared experiences with
public discourses, objects, situations, and events – such as
technology innovation. The invention of technical devices,
their diffusion in everyday life and the public discourse form
the generational consciousness and the understanding of tech-
nology. The experiences with technology that members of a
technology generation have during their youth and adolescence
form their attitudes towards and their way of handling tech-
nology. The technology they grew up with becomes a refer-
ence point and object of comparison. That’s why technology
generations vary e.g. in their competence using technology
(Sackmann & Weymann, 1994). This can be explained with
the concept of technology grammar (Rudinger, 1996), which
means knowledge about principles of technology usage as a
(cognitive) set of operation principles. Sackmann & Weymann
(1994) differentiate between four technology generations that
correspond with waves of technology innovation. Older adults
(age >55) belong to the “generation of increasing technization
of households” (born between 1949-1963) or to one of the
former generations (“generation of household revolution” born
1939-48; “pre-technical generation” born before 1939). They
grew up with home technologies (television, radio etc.) and
devices with mainly mechanical features. Operating with me-
chanical devices is radically different from operating with
electronic devices. Typical for using machines are e.g. press-
ing buttons with perceptible resistance, switches that have to
be turned over or sounds that indicate an action in progress. In
contrast, characteristically for electronic devices is the multi-
functionality and as a consequence several functions per key as
well as fully determined programme sequences that are not
perceptible for the user (Rudinger & Jansen, 2005). Lacks in
older adults technology grammar for electronic devices can
lead to operating errors, e.g. sometimes they use inadequate
operation principles. Because of the differences between the
technology grammar for mechanical devices and for electronic
devices it is hard to reason by analogy.
In order to get a deeper understanding in which ways the
technical performance of older adults is affected, an empirical
study was undertaken, in which older adults (age >55) and
younger (control group; age 23-29) were examined regarding
basic difficulties when using a software programme. The study
primary focused on visual aspects and language issues. The
aim of the present study was to identify pitfalls and shortcom-
ings older users experience and to clarify whether usability
problems are age-exclusive (causing interaction difficulties
only for older users) or age-specific (problems that cause seri-
ous interaction difficulties preferably for older people), and if
so, of what kind they are.
SAMPLE APPLICATION AND METHOD
The quality of a software programme named ‘Touch Speak
Designer’ was evaluated with a multi-method approach. The
software is part of the system Touch Speak. Touch Speak is a
computer based portable communication assistant for people
who suffer from a speech disorder, especially developed for
aphasia patients. Aphasia often occurs in older age as a conse-
quence of a stroke. Patients can use the communication aid to
complete sentences with words/phrases they cannot pronounce
or they can even substitute verbal communication with help of
the device (depending on the severity of their disorder).
The system consists of two components: a small screen de-
vice (communication aid) and the software programme to
configure the communication aid and to create and manage the
vocabulary. The software contains a database with about
10.000 so called speech-elements already available. Normally
the vocabulary and the device are configured by the attending
speech therapist or by family members. Approximately
two-thirds of the Touch Speak users are older than 60. The
probability that family members that have to deal with the
software are also older ones is relatively high. Thus the intui-
tive usability could be tested.
The software interface contains a toolbar and a menu bar
and is divided into two main areas: the database area (left) and
the area for the individual configuration of the vocabulary
(right) (see figure 1). Two additional areas (preview and simu-
lator) can be activated.
menu bar
toolbar
individual
database
vocabulary
Figure 1: Constituents of the sample applications interface
In a first step, the software was analyzed in a multiple cog-
nitive walkthrough to identify barriers and presumable usabil-
ity problems (expert evaluation). All available features were
examined taking the perspective of a typical user (old aged
family member that has to configure the communication aid).
The analysis based on general principles of web- and interface
design (Norman, 1988; Shneiderman, 2004; Wagner, 2002)
and specific ones for older adults (e.g. Coyne & Nielsen, 2002)
as well as principles for communication and interaction quality
(e.g. Grice, 1975). Furthermore, age-related functional deficits
and changes were considered as well as social factors (e.g.
technical socialization; Sackmann & Weymann 1994). The
result of the expert evaluation was a classification of problems,
which could lead to massive interaction difficulties for older
adults. Twenty different types of usability problems were iden-
tified and classified into five categories: Consistency, language
issues, feedback, layout and structure.
In a second step, older and younger participants were con-
fronted with the software programme in a usability test to
verify the results of the expert evaluation. Altogether, 14 vol-
unteers with different professional backgrounds participated (5
female, 9 male). Five of them were computer and/or software
experts; nine were classified as novices. This composition of
the sample should allow the identification of gender-specific
findings as well as the differentiation of age-specific problems
from those that were caused by lack of experience in working
with computers and various software programmes.
Participants had to solve six prototypical tasks, which were
embedded into a comprehensive scenario. The tasks were: (1)
identification of icons in the toolbox (naming), (2) orientation
(identification of the areas functions), (3) finding a speech
element in the database, (4) finding a specific function, (5)
creating a new speech element and (6) trying the simulation.
Every task had a time limit and varied in complexity and se-
verity: the participants started and finished with easy tasks;
task 3 and 5 were more complex (several steps to be taken) and
task 4 was in advance classified as rather difficult because the
function that had to be found was hidden in a context menu.
While solving the tasks the participants were instructed to
speak out loud what they were thinking, doing, and why they
were doing it (think aloud method). Their comments and
non-verbal behavior were recorded with video camera, screen
camera, and microphone. In a retrospective interview that
followed immediately, they were questioned about the tasks,
the usability of the software interface, and the test situation.
All collected data were transcribed, combined, and evaluated.
The analysis focused on three aspects: effectiveness (could
the tasks be solved within the time limit), time on task (how
long did they need to solve the tasks), and mainly qualitative
problems within the usage.
QUANTITATIVE RESULTS
Results show that the older users needed considerably more
time to complete the tasks and were less effective than the
younger ones. The more complex a task is the more distinct are
age differences. A comparison of task 1 (most easy task) and 5
(most difficult task) shows the differences: the younger par-
ticipants needed about 2 min on average to finish with task 1,
the older ones 3,4 minutes (see figure 2).
Figure 2: Time on task (in min.) – older vs. younger participants
The time difference for task 5 clearly exceeds (younger: 9,3
min; older: 14,3 min). The fastest participant (one of the
younger group) needed 16 min less than the slowest (one of
the older group). These results correspond with previous find-
ings (e.g. Baltes et al., 1995; Ziefle & Bay, 2005), according to
which task complexity led to a asymmetrically stronger per-
formance decrement in the older compared to the younger
group (Baltes, Lindenberger, & Staudinger, 1995).
Beside the time, the success rate was analyzed. Every sub-
task could be solved correctly (rated with 2 points), could be
partly solved (1pt) or not solved (0pt). A subtask was counted
as partly solved when the procedure was correct but realized at
a wrong time or place, e.g. when a speech element was in-
serted into the wrong category or a task was solved later on in
the testing. In this analysis, task 1 and 2 were left out because
the orientation and identification of icons (naming) could not
be evaluated as solved correct or wrong. Over all tasks, a total
score of 20 pt equals a 100% success rate) was to be reached.
While the younger adults reached a success rate between 80-90
% the older users, in contrast, showed a considerably lower
effectiveness (between 25-30 %). Only one of the older group,
a computer expert, was able to solve 85% of the tasks.
 The latter result may give a hint for an enormous potential
of compensating abilities (high expertise in combination with a
high variability within cognitive processing) as well in the
older group (Morrow et al., 1999).

QUALITATIVE RESULTS

In the following, qualitative findings are reported. The find- inconsistent

ings do correspond to the usability problem categories, which colouring
were identified in the expert evaluation. In order to visualize
the particular problem types and categories, descriptions of the
usability problems are flanked by quotes from casual remarks
and statements of participants in the retrospective interview.
inconsistent heading
Data were anonymized. The notation is as follows: O stands
for old and Y for young participant. E is the abbreviation for
computer experts, N for novices. R stands for a quote from a
retrospective, and I is the abbreviation for the Interviewer. Figure 3: Inconsistency in colouring and the use of headings

Inconsistent use of labels/ headings. The use of labels and

Consistency
headings is inconsistent if they are not used for all areas or
objects with the same or similar function without a reason.
The first category consists of consistency problems of dif-
Participants had difficulties in identifying the area to insert
ferent sources: inconsistent colouring, use of labels and head-
an appropriate picture – caused by an inconsistent layout re-
ings, terminology, and operations. Consistency means the
spectively a missing heading (see figure 3). The areas for the
consistent design of the interface itself as well as the consis-
configuration of title and message were in contrast clearly
tency to similar systems (under consideration of conventional-
headed. Thus many participants, primarily the older ones, did
ized design/patterns). Consistency is one of the most important
not notice the picture area as such and searched for a picture
principles of interactive systems. It is fundamental for the user
everywhere but not in the dialogue window. Three out of four
to understand the programme and to build up an adequate
older users could not solve the subtask to add a picture to the
conceptual model. A users conceptual model of a technical
new created element. Only one of the younger ones didn’t
system (software programme) bases on experiences with the
succeed. Again, we see that inconsistencies are bothersome for
system or similar systems (knowledge about operational rou-
all users, however, they interfere especially older adults inter-
tines and conventionalized patterns) and particularly on the
acting with the interface.
interface (Norman, 1988). If the software interface is designed
Inconsistent operations. If an identical operation is not re-
inconsistently users are confused, have difficulties to build up
peatable the same way, it is a matter of inconsistent operations.
a correct conceptual model, cannot understand how the system
For example when users want to configure the synthesized
works, get frustrated and, in the last resort, are not able to
voice output for a speech element sometimes the already in-
perform the required operation or task (Wagner, 2002). Fur-
serted message text is taken over automatically – sometimes it
thermore, consistency brings cognitive relief, facilitates auto-
is not and it has to be entered into a new inserting field again.
mation and learnability of operating processes, allows
This inconsistency makes the system unpredictable and leads
transferring acquired knowledge to unknown applications, and
to massive irritation. Two participants (YE 01 and OL 01)
helps to locate relevant information (Nielsen, 1993;
recognized this problem during the user test. As a consequence
Shneiderman, 2004).
the older participant (OL 01) did not solve the task because of
Inconsistent colouring. Inconsistent colouring means the
this problem; the younger participants (YE 01) on the contrary
usage of different colours for areas or objects with the same
sees through the problem after a while and conducts the opera-
function. Task 5 demanded that participants create a new
tion correctly.
speech element with different components (title, message,
Although inconsistent colouring of input fields and incon-
sound, picture). To do so, a dialogue window has to be opened
sistent labeling lead to a general irritation for users of both age
which contains different areas to insert and configure the
groups, they have a higher impact on the older users causing
components. Two of the older users could not solve this task
severe difficulties. Thus, consistency, particularly consistent
because of the inconsistent colouring of the input fields. Usu-
use of headings, labels and layout, is more important for older
ally input fields – in TouchSpeak Designer as well as in other
users than for younger ones. The conclusion to be drawn is,
software interfaces – are white. The input fields in the opened
that problems of consistency tend to be age-specific.
dialogue window for inserting the title or the message text are
grey (see figure 3). This colouring irritated the users as one
Language issues
stated in the retrospective interview: “Above, I inserted the
title because the curser was positioned there. But beneath, this
The category language issues includes all problems with
dark message window looked like it was disabled and as if you
language aspects of a software interface, e.g. naming of func-
can’t insert anything. That’s [grey colour for disabled] usual.
tions (menus, buttons), tool tips, and headings. In the expert
(…)” (RYE 01). This inconsistency did hinder the older par-
evaluation four different types were identified: general naming
ticipants more strongly than the younger, even though all par-
problem, reference problem, polysemy, use of technical terms
ticipants stumbled on this inconsitency.
and foreign words.
General naming problem. This category describes the false
or vague denomination of functions or objects (Jakobs &
Villinger, 1999). In the tool bar of the sample application, e.g.
is a button named “download”. Download means to receive
data to a local system from an external source. Contrary to this
common sense, in TouchSpeak Designer it denominates the
function to transfer data from the computer to the small screen
device, which is in fact an upload. This general naming prob-
lem caused irritation and misinterpretation: participants used to
interprete the button as “download” or “update from the Inter-
net”. As differences between the age groups could not be de-
tected, this type of problem is classified as not age-specific.
Reference problem. A reference problem occurs when there
is more than one possible reference object and neither the term
nor the context reveals the real reference object. In Touch-
Speak Designer, e.g. there is a menu item “editor for tool bars”.
However, there are several tool bars on the software interface
as well as on the communication aid to which this menu item
could refer. Similarly, several reference problems occurred
because of an inadequate separation of functions dealing with
the configuration of the communication aid and functions to
operate within the software. “Installation” or “hardware con-
figuration” for example are functions for the small screen
device, but are partly interpreted as functions to configure the
computer or the software, as the following statement from OE
02 shows: “Probably [a function] to configure the programme
according to individual preferences of the developer or user.”
Polysemy. Polysemy in contrast means the usage of one
term with different meanings and concepts to denominate
different objects or functions. Using polysem terms is prob-
lematic if ambiguity cannot be resolved by the functional and
linguistic context (e.g. as part of a specific menu). In the sam-
ple application the term “view” is a menu item to choose be-
tween different display options for the simulator (something
like a preview), but it is also used as a button to open an edit-
ing window for the configuration of sounds that belong to a
speech element. This leads to misinterpretations: Users do not
unequivocally identify the button as a possibility to open an
editing window; they mistake it is a preview for the inserted
text, as the following sequence shows:
I: “What did you think this button [view] is supposed
to be?”
OE 02: “That I can see the text.” (ROE 02)
Use of technical terms and foreign words. The use of tech-
nical terms or foreign words (Anglicism) can be problematic if
the user has not enough background knowledge about the
domain, respectively computers, and/or insufficient command
of English. The TouchSpeak Designer interface contains some
technical terms as well as Anglicism.
For example one button in the tool bar is labeled “client”,
which was incomprehensible for some of the younger users. “I
cannot image anything [what this could be]“ (RYE 02). All
older users identified the button correctly as “user administra-
tion”. Obviously this term is more familiar to the olds. A client
is a recipient of services, usually known from judiciary but
also used in the medical context. Probably the older ones have
more experiences with doctors and are more familiar with the
medical domain and domain specific terminology. We assume
that problems with domain specific terms decrease with in-
creased domain knowledge. In contrast computer specific
terminology causes interaction problems only for computer
novices – for younger as well as for older adults.
Summing up, results show no age differences regarding
this problem category. Obviously problematic language issues
lower usability for all users to the same extent. This result
corresponds with findings from the literature: suboptimal
names for functions have negative effects independent of the
users age (Schröder & Ziefle, 2005).
Structure
This category includes general structural problems, prob-
lems of categorization.
General structural problem. A general structural problem is
the use of an inadequate structural principle. According to
Norman (1991) it can be distinguished between eight structural
principles for interface elements, e.g. chronological order,
alphabetical order, semantic order. Depending on the purpose
of use, context and user different structural principles can be
adequate or not (K. L Norman 1991; Wagner 2002). The tool
bar in the examined software interface is structured by fre-
quency of usage: it contains the most often used functions
(according to the manual). But the order is designed according
to the mental model and usage of speech therapists, not to the
usage of people who use the communication aid at home. The
effects of this structural problem can be observed in the test as
a whole (not during a specific task). The participants mostly
search for required functions in the tool bar to solve the tasks.
They click on buttons like “client” and “hardware configura-
tion” that are irrelevant for their operations. The availability in
the tool bar signalizes “this function is important” and thus
they are delved into distraction. Other functions like “creating
a new element”, which are very important to solve the tasks –
and of course to work with the software at home – are not
present on the interface and can only hardly be detected (e.g.
in context menus). This user specific structure impedes the
usage of the software for people who use and configure the
system at home. The usage of the structural principle fre-
quency of usage is not adequate if the perspectives of different
user groups diverge like in this example.
Problems of categorization and assignment. In the expert
evaluation several problems of categorization and assignment
were identified. Both problem types are strongly related: any
assignment problem is as well a problem of categorization. A
problem of categorization in this case is a wrong categoriza-
tion in terms of the denomination of a group of functions.
Problems of categorization occur mainly in the menu bar.
Previously one categorization problem was estimated to be
severe: the function “creating a new speech element” is located
in a context menu and can only be found if users open the
context menu via right click in the vocabulary area (right area).
This positioning violates the principle of visibility (Norman,
1988). Elements must be eye catching in order to attract users
attention, beyond the demands of a good visual ergonomic
design. The importance of this principle can be shown when
considering participants’ difficulties in task 4. Here, the func-
tion “creating a new speech element” had to be detected. The
result was clear: none of the participants could find the func-
tion. They searched nearly everywhere, consulting the online
help and trying various strategies to find the function and to
solve the task. Most of the users assumed the function to be
located in the database area. YE 02 explains in the retrospec-
tive interview, why he/she searched in this area: “I thought, on
the left side, that there is the database and everything that
already exists. I want to extend this [data set]. Probably I want
to reuse it [new speech element]. And if you only create it [in
the vocabulary area], it is only on the one device and not left
[database].” (RYE 02). This explanation shows that the posi-
tioning of the function is not transparent and intuitively under-
standably, but rather causes severe interaction problems. As a
consequence, users are hindered to solve the task and, in the
last resort, are prevented to use the communication aid effec-
tively. This problem is not age specific, neither older users nor
younger ones could find this function. Schröder & Ziefle (2005)
come to corresponding results: They tested menu structures of
mobile phones with younger and older participants and found
out that poor function categorization and assignment lead to
interaction difficulties independent of the age.
Altogether no significant age differences were identified for
the category structural problems.
Feedback
Feedback is the response (of a system) to the user. It shows
which action has been conducted and which result has been
achieved (Norman 1989). Feedback can be given in various
forms: visual (change of an object or a status indication),
acoustic (sound, alert) or textual via message or dialogue win-
dows or in the status line of applications (Wagner, 2002). In
this case problem reports, error messages and security queries
belong to the feedback forms as well.
The expert evaluation led to four different types of feedback
problems: wrong feedback, non-informative feedback, lack of
feedback and imperceptible feedback.
A security query or verbal feedback is reasonable if a task
comprises several steps in succession. Complex operations
offer more potential spurious actions and identifying sources
of errors is difficult. In task 5, creating a new speech element
with title, message, voice output and picture, included such
complex subtasks and several feedback problems, which
caused severe interaction difficulties.
To activate the speech output for example five operating
steps are necessary (see figure 4).
1 indication]. That is not user friendly. At least one could show a
Play 2 warning that this way [with this configuration] nothing is out-
3 output is activated is the enabled play button in the dialogue
5 window “editor for special speech elements”, which can be
4
Figure 4: Operational steps configurating voice output
Because this example is rather complex, it will be explained
in detail: the speech output for title and message can be edited
in the dialogue window “editor for special speech elements”.
Participants should configure the speech element in such a way
that the message is verbalized by a synthesized voice (two
voice options exist: natural voice and synthesized voice).
Therefor, a checkbox named “sound” beyond the inserting
field for the message has to be activated first. By activating the
checkbox, the button “view” next to it is enabled, indicated by
the change of the font colour from grey to black (visual feed-
back). Then the user has to click the button “view”. A new
dialogue window “sound studio” opens. In this dialogue win-
dow the checkbox “using synthesized voice” has to be acti-
vated. By doing so the upper area for editing existing
digitalized sounds is disabled (visualized by a change of the
button and type colour from black to grey) and the lower area
for the voice output with a synthesized voice is enabled (visu-
alized by the colour change of the button “speak” from black
to grey). The fourth step to be conducted is entering the text
into the inserting field next to the button “speak”. The action
has to be finished by clicking the button “ok”. When the
checkbox “using synthesized voice” has been activated and the
dialogue window “sound studio” has been closed, a button to
play the sound is enabled (indicated by a visual feedback:
colour change from grey to blue).
In this subtask various problems interfered. Some partici-
pants did not even open the dialogue window “sound studio”.
But those who did were confronted with the problem types
lack of feedback and imperceptible feedback. The steps de-
scribed above have to be conducted without any further infor-
mation or instruction. Most of the participants who reached the
dialogue window, activated the checkbox “using synthesized
voice”, but then they did not recognize that they have to enter
the message text into the designated inserting field. The mere
existence of an inserting field is not indication enough to
evoke participants’ attention and to bring them to enter text
into it. An additional feedback, in form of a message or an-
other hint is necessary. The next problem occurred immedi-
ately after, when users wanted to finish the configuration of
the voice output: user closed the window via the “ok”-button,
but did not get a feedback whether their action was successful
or not. Irritated by this, some users opened the window again
and wanted to check whether the voice output is correctly
activated or not, but they did not exactly know how to find out.
One participant criticizes this lack of feedback: “Here, in the
“view”, I have to comment this again, there is nothing [no
putted.” (YE 01).
Even for those who were successful in activating the voice
output with the correct message imperceptible feedback turned
out to be problematic. Imperceptible feedback is a feedback
that can hardly be seen or perceived by users because it hap-
pens outside his focus of attention and/or is too small or unre-
markable. To explain this type the above mentioned example is
picked up once again: the only way to ascertain that the voice
seen if the “sound studio” is closed.
But the colour change of the button is nearly imperceptible
for users because the time span between seeing the original
state (disabled button coloured grey before opening the “sound
studio”) and the new state (enabled button coloured blue after
activating the sound) is too long and the object and its change
are too small. This imperceptible feedback problem made users
insecure and they wasted a lot of time, because they were
looking for a sign that the action was successful.
Within the same operation another feedback problem was
detected: the play button sometimes also changes as well when
no sound has been activated. This is obviously a bug. But in an
interaction progress it must be evaluated as a wrong feedback
given to the user. The effects of this problem type can not be
estimated based on the user test because they are concealed by
the first mentioned problem of imperceptible feedback: as
participants do not notice the enabled play button as confirma-
tion of an activated sound, they do also not recognize that it
shows an available sound if there is none.
The problem type non-informative feedback is defined as
feedback or messages that inform the user about the result of
an action, but the contained information is incomplete and/or
irrelevant and thus does not help the user to proceed in interac-
tion. This problem type occurred for example in task 3. Par-
ticipants had to search an existing speech element
(“strawberry”) in the database. In advance, several correct
approaches to search for elements were identified: the first and
fastest is the search function in the database area. Furthermore,
a separate dialogue window can be opened via the button
“search” in the toolbar or the menu item “search” in the tool
menu of the menu bar. OL 06 uses the database search, but
enters the word “strawberry” unintended with leading blanks.
Thus the system can not identify the query and responds with
the message “TouchSpeak Designer” did not find a matching
object.” This message is correct, but does not help the user to
identify his typing mistake. It does not tell the user why no
matching objects could be found. He tries various spellings
(e.g. with quotation marks), but in all his attempts, he does not
remove the leading blanks. OL 06 comes to the conclusion that
the operation he conducted is incorrect: “I don’t know if the
quotation marks belong to the entry, or if it is only ‘straw-
berry’ [without quotation marks that has to be entered]. But
probably, I am on the wrong track.”
In this case the user is handicapped because of his limited
sighting ability. He does not recognize his typing mistake and
is not able to identify it on his own. The system does not sup-
port him. As the statement above shows, this usability problem
does not only lead to the users’ failure in solving the task, but
also to a loss of trust in his own competence. This behaviour is
typical for older users (e.g. Göbel & Woo Yoo, 2005).
One of the younger participants makes the same mistake
during another task, but he has no problems in detecting and
correcting it. Frequently, it was observed that older users tend
to make more often small mistakes, which they though do not
recognize and which disturb the interaction.
Obviously, older adults need more often a correct, informa-
tive, and perceivable feedback than younger adults do – par-
ticularly they need support in trouble shooting. Younger
people profit from good feedback as well. But they are rather
able to locate, and eliminate mistakes on their own, and to
cope with cognitively complex tasks without detailed instruc-
tions. Altogether this problem category tends to be
age-specific.
Layout
The last category consists of five problem types: font prob-
lems, poor colouring and contrasts, problems of operational
elements, inappropriate location, and disturbed relationship
between form and function.
Font problems. This type subsumes problems of font type
and size. It is recommended to use a sans-serif font type and at
least 12 pt to 14 pt type size, 16 pt for headings (Zhao, 2001).
Furthermore Coyne/Nielsen (2002) suggest for websites to
offer a button to increase the type size for older people with
visual handicaps (Coyne & Nielsen, 2002). In TouchSpeak a
sans serif font type is used but headings size and button labels
are too small. A possibility to increase font sizes is not avail-
able. This problem occurs in connection with a poor colouring
respectively contrast.
Poor colouring and contrast. Poor colouring and contrasts
are more serious for older adults due to limited sensory abili-
ties (Kline & Scialfa, 1997). Therefore it is recommended to
“maximize the contrast between foreground and background
colors: use dark types on light or white backgrounds, exagger-
ate lightness differences between foreground and background
colors, and avoid using colors of similar lightness adjacent to
one another.” (Zhao, 2001). Looking at the headings in the
sample application it becomes clear that they are designed
suboptimally: white font on blue background, which can
“cause the type to appear to ‘close’ in itself (Zhao, 2001). Also,
too small font types and poor colouring (e.g. in task 2) hinder
users to build up an appropriate mental model of the pro-
gramme: only one (!) out of 14 participants did recognize the
headings of the main areas.
Operational elements. This type summarizes problems of
size, type ad labeling of operational elements. Because of
functional deficits in fine motor skills and shiver older people
often have difficulties in controlling the mouse (movements,
clicking, dragging) and positioning the cursor pinpoint (Czaja
& Lee, 2003). For this reason small operational elements
should be increased, and scrollbars should be avoided as well
as pull down menus. Between operational elements should be
enough space so that older users do not accidentally click on
the wrong element. “The larger the target and the more space
between targets, the more likely it is that seniors will hit it on
the first attempt” (Coyne & Nielsen, 2002). Keeping this in
mind, it is obvious that the buttons in the database and in the
individual vocabulary area do not meet the requirements of
older users: they are undersized and too close to one another
see figure 5).
1 2 3 4 5 6 7 8
Figure 5: Operational elements in the database area
Although participant OL 06 previously had changed the
screen resolution (because of visual deficits he normally uses
additional software to increase the screen display) he could
hardly perceive the buttons. He describes his difficulties in the
retrospective interview: “I have a programme and I help my-
self by changing to 800, 600 [dpi]. That is not as sharp as 1000.
Yes, this, the labels [tooltips] are easy to see, but those [but-
tons] are curious.”
 Furthermore, the buttons are not labeled. Only tooltips are
displayed on mouse over and show their meaning. Icons and
buttons should be labeled if they are ambiguous or they are
unknown to some users (Horton, 1994). Irritation arose be-
cause well-known symbols (taken from everyday life) were
used for programme specific elements, and, accordingly, could
not be correctly interpreted as the following quote shows:
“Left, I cannot imagine what this is. Second is some document.
Third is television. Fourth is communication.” (OL 06)
The first four buttons are filter options for the database (see
figure 5). The first icon (a book) stands for the function “show
only general speech elements”. The proband cannot even iden-
tify the book probably because of its size – even less he knows
what it means. The second (a document) represents “show only
special speech elements”, and is interpreted as “some docu-
ment”. The third (an easel) “show only pictures”, reminds the
user of a television and the fourth (a stylized organigram, also
used in other areas of the interface) means “show only vo-
cabularies”. The user characterizes it as “communication”. He
misinterprets the symbols despite of the explicative tooltips.
Obviously they do not help him because he does not know the
displayed terms. They are specific for the programme. Neither
the labels nor the symbols/icons can activate correct mental
concepts of the functions. In this case, the buttons should be
clearly labeled with comprehensible terms to ensure that users
can work effectively with the database area.
Test results show that the undersized operational elements
are problematic for older users. Although missing labels
(headings and labels of buttons/input fields) irritate younger
users as well, they lead to severe problems for the elderly.
Inappropriate localization. This problem type includes both,
absolute and relative positioning misunderstandings. Absolute
localization positions of interface elements are suboptimal if
users have difficulties in perceiving them because they are
outside their attention focus. In western cultures the common
reading direction is from left to right and from top to bottom.
Accordingly, users’ attention focuses on the upper left quarter
of a screen display. Other screen localization of objects pro-
vokes suboptimal information procession. In the evaluated
software interface this ergonomic principle is violated (e.g. the
inadequate spatial positioning of the scrollbar for the database
area in list view; see figure 6).
Figure 6: Inadequate positioning of undersized objects
Because of its misleading position and size many users do
not recognize this operational element. Again, this shortcom-
ing mainly affects older users navigation behaviour, while
younger users did recognize the scrollbar, but criticize the
small size of this element. One younger user characterizes this
in the retrospective interview very clearly: “Well, I don’t know
but perhaps in this view here, probably these are too small.
Probably if one is older it is difficult to read. The left area here,
the three displaying options are very small and at first one
doesn’t see it (…).” (RYE 02)
Obviously, all users recognize layout problems, but they only
cause troubles for older ones. The same effect was revealed
with relative localization problems. Spatial proximity of ob-
jects shows togetherness (law of proximity) (Goldstein, 1997).
According to the law of proximity, objects (functions, buttons,
headings, labels) that belong together should be bundled. In
the search window that can be opened via toolbar or menu bar
this law is violated: the label of the tab for search and the
search field are separated. The distance between them is too
big to indicate their togetherness and no other cues show the
function of the field directly. This can be demonstrated in the
next interview extract:
I: Why didn’t you recognize it [the search field]? Did
you miss anything? Is there anything you required at
this point?
OL 05: I actually didn’t read this “search”. I didn’t perceive
it.
Coyne and Nielsen (2002) come to similar findings. In their
web usability study for seniors, older adults had difficulties to
identify a search field as such when it is not clearly labeled. As
none of the younger participants in our study had comparable
difficulties, the relative localization problem can be classified
as an age specific problem.
Weak relationship between form and function. According to
the design principle “form follows function” respectively
“function follows form”, the form or design of things should
be derived from its function and vice versa. This principle can
also be applied to interface design: functions marked by a
special visual appearance (form, colour) follow this principle
and support an intuitive usage (Chisnell, Redish, & Lee, 2006).
The same ergonomic benefit is given in conventionalized rela-
tionships, which have been established by experience and
learnability effects to a standard over years of usage. Enabled
input fields, for example, are usually white coloured, and have
a three dimensional visual appearance. In the analyzed inter-
face two violations of this convention have been detected: in
the dialogue window “editor for special speech elements”
input fields have a grey background though they are enabled.
In the database area a large field with a white background was
presented although it is no input field. By this colouring the
conventionalized relationship between form (white) and func-
tion (input field) is disturbed. This caused irritation during the
user test: in task 2, participants had to determine which func-
tions the different areas of the software interface have. Most of
them misinterpreted the database area as an editor, where new
speech elements can be entered, as the following remark clari-
fies: “In area one, I think that it is a kind of editing field in
which I can describe such an area of life and link it with com-
mands, instructions, questions, and so on.” (OE 02)
Layout problems and ambiguities between form and func-
tion not only cause irritation, but also result in a mismatch
between users’ mental model and designer’s conceptual model.
Users cannot build up an appropriate mental model of the
system. This problem type impedes older users more than
younger ones.
In summary the category layout problems shows a distinct
tendency to be age specific. The problem types of this category
mainly lead to difficulties for older users, while younger adults
succeeded to compensate the ergonomic shortcomings.
DISCUSSION
In this study, we examined usability problems of older users
when using specific software for aged users. Participants were
asked to solve typical tasks, and were observed regarding basic
difficulties and shortcomings of the interface design. In order
to evaluate the extent of age-related problems, a control group
of younger adults were also examined. Main research aims
were the identification of problems, which are mainly
age-exclusive (causing interaction difficulties only for older
users) or age-specific (problems that cause serious interaction
difficulties preferably for older people).
Interestingly enough, none of the identified problems turned
out to be age-exclusive. The shortcomings of the interface
caused irritation and interaction difficulties for both groups.
This especially regards language and structural problems,
which were problematic for both age groups and decrease
usability independently from user characteristics. More
age-specific shortcomings are caused by inconsistency, unclear
feedback, or layout problems, which occurred more often for
older users, causing severe troubles and perceived helpless-
ness.
Older users especially failed whenever they had to solve
complex tasks in combination with a suboptimal designed
feedback. Complex tasks, comprising many steps in succession,
offered more room for mistaking. Obviously, older users need
more (linguistic) support (qualitative and quantitative) to iden-
tify and to correct mistakes. If there is no appropriate feedback,
they cannot complete complex tasks successfully on their own.
This result can be explained with sensorial and cognitive
changes in older age: because of a declining perception (visual
and auditory changes) in older ages it could be harder for older
people to detect a visual or acoustic feedback particularly if the
feedback is only a minimal change of an object or signal. For
older people feedback should be designed more obvious, per-
ceivable and permanent (Hawthorn, 2000; Wagner, 2002). In
addition, linguistic messages could be conducive to the cogni-
tive relief when complex operational sequences are divided
into sub-steps and all sub-steps finish with a feedback message
(security query). Thus the amount of information that has to be
processed at the same time is lowered. The likelihood that
mistakes are immediately detected is higher, and the user has
more control over the ongoing processes.
Difficulties caused by inconsistency tend to be age-specific.
This is reasonable because an inconsistent design increases the
cognitive load, impedes automation and learnability of operat-
ing processes, as well as transferring acquired knowledge to
unknown applications, and the location of relevant information
(Nielsen, 1993; Shneiderman, 2004). Because of the cognitive
changes in older age (e.g. longer information processing time,
impeded information search, difficulties in solving novel
problems and localizing objects) a consistent design is even
more important for older than for younger users. In addition
consistency is the most important precondition for under-
standing a technical system. Older users want to understand
the technologies they use and therefore an inconsistent design
that makes a technical system hard to understand hits older
users more strongly than younger users (Jakobs, et al., 2008;
Ziefle & Bay, 2004).
Focusing only on the language aspects of software interfaces
it becomes clear that primarily a lack of linguistic elements
(lack of function names, headings, feedback, linguistic hints or
operational instructions) impedes the interaction for older
users. It has to be proved if an increased usage of linguistic
elements can improve software interfaces for older adults.
Interestingly, older people tend to be more familiar with the
medical domain and domain specific terminology whereas
younger users are not. This result points out that the domain or
using context and the users experiences with this domain have
to be considered for the design of interfaces as well. This result
could for example be relevant for the choice of metaphors,
scenarios, and operational sequences of future applications.
One could argue that the problems the older adult group ex-
perienced are mainly due to the specific software which was
under study. Even if this objection cannot be fully ruled out
many of the findings reported here represent universal short-
comings in interface designs, which can be observed in many
software applications and interface types, independently of the
specific content and software environment. Based on this the
findings and the conclusion can be classified as transferrable to
other software applications.
CONCLUSION
The conclusion to be drawn is that the visual and linguistic
design of software interfaces has to be carefully considered to
meet the requirements of older users. Age-friendly designs are
not restricted to visibility and readability issues (font sizes,
color or contrast). Still more important are the design of the
interaction with the device, the design of (verbal) feedback and
the usage of detailed operational instructions. Further studies
will have to show if the need for detailed communication de-
pends on the type of application, the type of device, the type of
technology (e.g. information and communication technology),
the domain or using context as well as the complexity of the
interface. Overall, for an age-sensitive design it is indispensi-
ble to rely on the existing knowledge of different disciplines
and to pursue a truly multidisciplinary approach. This
multidisciplinarity is the only way to fully understand the
aging problem, to sensitively describe and examine aged users
when using technology, and, to develop age-sensitive con-
cepts.
REFERENCES
Baltes, P., Lindenberger, U., & Staudinger, U. M. (1995). Die
zwei Gesichter der Intelligenz im Alter [The two faces of
intelligence in old age]. Spektrum der Wissenschaft, 10,
52-61.
Baltes, P. B. (1991). The many faces of human aging: Toward
a psychological culture of old age. Psychological Medicine,
21, (pp. 837-854).
Chisnell, D. E., Redish, J. C., & Lee, A. (2006). New
Heuristics for Understanding Older Adults as Web Users.
Technical Communication, 53(1), 39-59.
Coyne, K., & Nielsen, J. (2002). Web Usability for Senior
Citizens: Design Guidelines based on Usability Studies with
People Age 65 and older Fremont: Nielsen Norman Group.
Czaja, S. J., & Lee, C. C. (2003). Designing Computer
Systems for Older Adults. In J. J. Jacko & A. Sears (Eds.),
The Human-Computer Interaction Handbook. Fundamentals,
Evolving Technologies and Emerging Applications (pp.
413-427). New Jersey/ London: Erlbaum
destatis (2008). Statistisches Jahrbuch 2008 für die
Bundesrepublik Deutschland. Wiesbaden: Statistisches
Bundesamt Deutschland [Statistical Yearbook for the
Federal Republic of Germany 2008]
Göbel, M., & Woo Yoo, J. (2005). Anforderungen älterer
Menschen an Moderne Technik und deren Dokumentation.
In C. Schwender (Ed.), Technische Dokumentation für
Senioren (Vol. 12, pp. 63-82). Lübeck Schmidt-Röhmhild
[Older adults requirements for modern technologies and
their documentation]
Goldstein, E. B. (1997 ). Psychology of perception. Berlin:
Spektrum Akademischer Verlag.
Grice, H. P. (1975). Logic and conversation Syntax and
Semantics, 3, 41–58.
Hawthorn, D. (2000). Possible implications of aging for
interface designers. Interacting with Computers, 12(5),
507-528.
Horton, W. (1994). The icon book: Visual symbols for
computer systems and documentation: New York: Wiley &
Sons.
Jakobs, E.-M., Lehnen, K., & Ziefle, M. (2008). Alter und
Technik. [Age and Technology]. Aachen, Germany:
Apprimus
Jakobs, E.-M., & Villinger, C. (1999). „Das versteht kein
Mensch...“. Verständliche Gestaltung von Hilfesystemen für
Softwareprogramme ["Nobody unterstands this…".
Comprehensible design of help systems for software
programmes.] In E.-M. Jakobs, D. Knorr & K.-H. Pogner
(Eds.), Textproduktion. Text, HyperText, KonText (pp.
217-233 ). Frankfurt, Germany: Lang.
Kline, D. W., & Scialfa, C. T. (1997). Sensory and Perceptual
Functioning. In A. D. Fisk & W. A. Rogers (Eds.), Human
factors and the older adult. (pp. 27-54). San Diego:
Academic.
Morrow, D., Miller, L.S., Ridolfo, H., Kokayeff, N., Chang, D.,
Fischer, U. and Stine-Morrow, E. (2004). Expertise and
aging in pilot decision-making task. In Proceedings of the
Human Factors and Ergonomics Society 48th Annual
Meeting, Santa Monica, CA. (pp. 228-232).
Nielsen, J. (1993). Usability engineering: Morgan Kaufmann.
Norman, D. (1988). The psychology of everyday things: Basic
books New York.
Norman, K. L. (1991). The psychology of menu selection:
designing cognitive control ot the human/computer interface.
Norwood: Ablex Publishing Corporation.
Oetjen, S. & Ziefle, M. (2007). The Effects of LCD’s
Anisotropy on the Visual Performance of Users of Different
Ages. Human Factors, 49(4), 619-627
Park, D.C. & Schwarz, N. (1999). Cognitive Aging.
Phildelphia: Buchanan.
Röhr-Sendlmeier, U. (1993). Kompetenz im Alter: Der Beitrag
der Technik. [Competence in the every day life: the
contribution of technology]. Paedagogische Rundschau,
47(1), 69-83.
Rudinger, G. (1996). Alter und Technik. [Age and Technology]
Zeitschrift für Gerontologie und Geriatrie, 29, 246-256.
Rudinger, G., & Jansen, E. (2005). Technik, Neue Medien und
Verkehr [Technology, new media and traffic]. In S.-H.
Filipp & U. M. Staudinger (Eds.), Entwicklungspsychologie
des mittleren und höheren Erwachsenenalters. Göttingen,
Germany: Hogrefe.
Sackmann, R., & Weymann, A. (1994). Die Technisierung des
Alltags. Generationen und technische Innovationen [The
technization of the every day life. Generations and technical
innovations]. Franfurt, Germany: Campus.
Schlick, C.; Ziefle, M.; Park, M. & Luczak, H. (2007). Visual
displays. In J. Jacko and A. Sears (eds.). The Human
Computer Interaction Handbook: Fundamentals, Evolving
Technologies and Emerging Applications. Chapter 11 (pp.
201-228). Mahwah, New Jersey: LEA.
Schneider, N., Schreiber, S., Wilkes, J., Grandt, M., & Schlick,
C. (2008). Foundations of an age-differentiated adaptation of
the human-computer-interface. Behaviour & Information
Technology, 27(4), 319-324.
Schröder, S., & Ziefle, M. (2005). Semantic transparency of
ellular phone menus. Comparing users from different age
groups. Computer Studies in Language and Speech (pp.
302-315). Frankfurt, Germany: Lang.
Shneiderman, B. (2004). Designing the User Interface:
Strategies for Effective Human-Computer Interaction
Amsterdam: Addison-Wesley.
Vercryssen, M. (1997). Movement control and speed of
behavior. In A. D. Fisk and W. A. Rogers (Eds). Human
factors and the older adult (pp. 55-86). San Diego:
Academic.
Wagner, J. (2002). Mensch-Computer-Interaktion: sprach-
wissenschaftliche Aspekte: [Human-Computer- Interaction:
linguistic aspects]. Frankfurt, Germany: Lang.
Zajicek, M. (2001). Interface Design for Older Adults. Paper
presented at the EC/NSF workshop on Universal
accessibility of ubiquitous computing: providing for the
elderly.
Zhao, H. (2001). Universal Usability Web Design Guidelines
for the Elderly (Age 65 and Older).
www.otal.umd.edu/uupractice/elderly/
Ziefle, M. & Bay, S. (2004). Mental models of Cellular Phones
Menu. Comparing older and younger novice users. In S.
Brewster & M. Dunlop (eds.). Mobile Human Computer In-
teraction (pp. 25-37). Berlin: Springer.
Ziefle, M. & Bay, S. (2005). How older adults meet cognitive
complexity: Aging effects on the usability of different
cellular phones. Behaviour &Information Technology, 24(5),
375-389.
Ziefle, M, & Bay, S. (2006). How to overcome disorientation
in mobile phone menus: A comparison of two different types
of navigation aids. Human Computer Interaction, 21(4),
393-432.
Ziefle, M. & Bay, S. (2008). Transgenerational Designs
in Mobile Technology. In J. Lumsden (Ed.). Handbook of
Research on User Interface Design and Evaluation for
Mobile Technology (pp. 122-140). IGI Global.